Are Breast Implants Better than Female Genital Mutilation? Autonomy, Gender Equality and Nussbaum s Political Liberalism

Transcription

1 1 Are Breast Implants Better than Female Genital Mutilation? Autonomy, Gender Equality and Nussbaum s Political Liberalism CLARE CHAMBERS This essay considers the tension between political liberalism and gender equality in the light of social construction and multiculturalism. The tension is exemplified by the work of Martha Nussbaum, who tries to reconcile a belief in the universality of certain liberal values such as gender equality with a political liberal tolerance for cultural practices that violate gender equality. The essay distinguishes between first- and second-order conceptions of autonomy, and shows that political liberals mistakenly prioritise secondorder autonomy. This prioritisation leads political liberals to seek to limit state interference in individuals choices. However, the essay argues that if options, choices and the preferences which lead to them are socially influenced or constructed, it is no longer clear that state non-interference secures autonomy. Instead, it becomes a matter of justice what the content of the social or state influence is, which options are open to people, and political liberalism cannot deal with many forms of injustice. Rather than emphasising state neutrality, liberals should endorse state prohibition of practices which cause significant harm to those who choose them, if they are chosen only in response to unjust norms. FCRI7301.sgm / Critical Original 2004 Taylor Autumn ClareChambers Government ce.chambers@lse.ac.uk. and & Review Article Francis print/ DeptLondon of 2004 Ltd International Ltd School online Social of Economics and Political & Political Philosophy ScienceHoughton StreetLondonWCZA ZAE Why aren t liberals willing to liberate? Increasingly, contemporary liberals such as Brian Barry and Martha Nussbaum are happy to criticise oppressive practices, wherever in the world they occur. These liberals, however, stop short of advocating state action to lessen the oppression resulting from those practices, claiming that the state cannot intervene in matters of substantive morality. As a result, such liberals condemn many individuals to oppressive and non-autonomous lives. This essay seeks to Critical Review of International Social and Political Philosophy, Vol. 7, No. 3, Autumn 2004, pp ISSN print/ online DOI: / Taylor & Francis Ltd

2 2 CRISPP understand how liberal conceptions of universal autonomy can be so selective in their application, and to rectify that selectivity. In particular, it focuses on the problems caused by political liberalism s prioritisation of what I call second-order autonomy. Generally, political liberals want to protect individuals freedom of choice, even if it is used to choose things that are harmful to the choosing individual. For example, Rawls places great store by individuals ability to frame, revise and pursue their conception of the good. For Rawls, this sort of choice is so important that it underpins the lexical priority given to the equal basic liberty principle in A Theory of Justice. In Political Liberalism and The Law of Peoples Rawls continues and strengthens this emphasis on choice, entrenching individuals right to choose to live what we might otherwise think of as non-liberal ways of life. Thus in The Law of Peoples Rawls rejects the idea that comprehensive liberal principles of justice ought to apply globally, with a global difference principle regulating the international distribution of wealth and the equal basic liberties applying to all. Instead, he argues, non-liberal yet decent hierarchies must be tolerated by liberal states a position that is justified at least in part by reference to other peoples ability to choose different, non-liberal ways of life. I want to argue, though, that a truly liberal project must take account of two aspects of social construction. The first aspect of social construction is the way in which individuals and their preferences are formed by social forces, so that who people are and what they want are affected by their social context. The second aspect of social construction, and the one focused on in this paper, is the fact that individuals options are constrained by social norms some of which are harmful or epitomise inequality. As such, the options available to people, or the options which are most appropriate for people to choose, may have negative implications for the individual s equality and freedom. As a result of these two aspects of social construction, political liberalism is not well placed to secure liberal justice. In this essay, I extend my argument through analysis of the recent work of Martha Nussbaum. Nussbaum s approach is useful and interesting as it combines a commitment to Rawlsian political liberalism with a critique of harmful social norms and an awareness of social construction. As such, I take as my starting-point the notion that Nussbaum s work is an improvement on alternative politically liberal accounts, and that it will not be a solution to the problems I highlight to abandon her distinct contributions and return to a more standard Rawlsian approach. However, as I shall show, Nussbaum s political liberalism prevents her from making good the

3 AUTONOMY, GENDER AND POLITICAL LIBERALISM 3 promise of her work on social construction and the injustice of social norms. Much of what is valuable about the latter aspect of her work, moreover, undermines or directly contradicts her insistence on political liberalism. In general, political liberalism is peculiarly ill equipped to deal with injustices resulting from culture and choice as it abandons significant areas of justice to be determined by individuals choices. Autonomy and Political Liberalism To begin with, let me highlight the aspect of political liberalism that I think is most problematic in the light of the two aspects of social construction. Political liberalism, according to Nussbaum, is characterised by prioritisation of what I call second-order autonomy. Nussbaum contrasts her political liberalism with the comprehensive liberalism of John Stuart Mill, Joseph Raz and Susan Moller Okin. For these comprehensive liberals, Nussbaum argues, autonomy is a general good for all humans: liberal values of autonomy and dignity pervade the fabric of the body politic and the fostering of personal autonomy [is] an appropriate goal of the state (1999b: 108). Nussbaum argues that her position, in contrast, allows people to live nonautonomous lives, for autonomy may be counter to their conception of the good, particularly if that conception is religious. Thus the political liberal endorses the idea that each citizen as citizen is an equal chooser of ends, but she carefully refrains from asserting that nonautonomous lives are not worth leading (ibid.: 110). Nussbaum s argument on this point follows Rawls, who writes: Full autonomy is achieved by citizens: it is a political and not an ethical value. By that I mean that it is realized in public life by affirming the political principles of justice and enjoying the protections of the basic rights and liberties; it is also realized by participating in society s public affairs and sharing in its collective self-determination over time. This full autonomy of political life must be distinguished from the ethical values of autonomy and individuality, which may apply to the whole of life, both social and individual, as expressed by the comprehensive liberalisms of Kant and Mill. Justice as fairness affirms this contrast: it affirms political autonomy for all but leaves the weight of ethical autonomy to be decided by citizens severally in light of their comprehensive doctrines. (Rawls 1993: 77 8; 2001: 156).

4 4 CRISPP The question which then arises is whether a nonautonomous life must be autonomously chosen if it is to be acceptable to the political liberal. In order to clarify this question it is necessary to investigate the complex concept of autonomy further. Autonomy has been used in a great many different ways by different philosophers (see Dworkin 1988: 5 6 for some examples) and some distinctions are needed. For my purposes, it will be necessary to distinguish three concepts: negative freedom, and what I will call first- and second-order autonomy. The first concept, negative freedom, is familiar though contested. Its nuances are not crucial in this context; the basic idea of negative freedom is a lack of interference from others or, to be more specific, a lack of coercion (Berlin 1984; Hayek 1960). The second and third concepts I term two orders of autonomy. Second-order autonomy applies to the manner in which an individual comes to have a particular way of life or comprehensive conception of the good. One is second-order autonomous if one actively and willingly chooses one s way of life free from compulsion or influence which would obscure that choice. Thus Rawls defence of individuals moral power to have, to revise, and rationally to pursue a conception of the good (Rawls 2001: 19) can be thought of as a defence of second-order autonomy: individuals must be able to lead the way of life that they choose for themselves. First-order autonomy applies to one s attitude to the rules and norms which are part of a way of life. One is first-order autonomous if one leads a daily life in which one questions rules and norms and actively chooses how to respond to them. One may be autonomous and follow rules, but only if one considers the rule and decides that it is a good rule to follow. One is autonomous if one is governed by rules that one sets for oneself or endorses for oneself. This conception of autonomy has many followers, as Gerald Dworkin points out: As a moral notion shared by philosophers as divergent as Kant, Kierkegaard, Nietzsche, Royce, Hare, and Popper the argument is about the necessity or desirability of individuals choosing or willing or accepting their own moral code. We are all responsible for developing and criticizing our moral principles, and individual conscience must take precedence over authority and tradition (Dworkin 1988: 10 11). The distinction between first-order autonomy and negative liberty, though necessary for analytical clarity, can introduce confusion. Consider the question, for example, of whether life in a convent or army

5 AUTONOMY, GENDER AND POLITICAL LIBERALISM 5 FIGURE 1 FOUR SCENARIOS OF AUTONOMY is autonomous. Clearly, if it is chosen, army or convent life is compatible with second-order autonomy. Equally clearly, such lives entail great constraints on negative liberty. The position regarding first-order autonomy, however, is ambiguous. Depending on the nun or soldier s attitude to the rules of the institution, first-order autonomy may or may not be present. If she endorses all the rules, then she will obey them with firstorder autonomy, even though her liberty is constrained. However, on the reasonable assumption that the nun or soldier has to obey at least some rules which she finds problematic, both first-order autonomy and negative liberty are limited. This issue is important and will be returned to. For now, though, assume that at least some of the rules of a convent or army are not endorsed by its members, so that the lack of negative freedom about which rules to obey leads to a lack of first-order autonomy. With this assumption in mind, consider four distinct logically possible scenarios (Figure 1). FIGURE 1 FOUR SCENARIOS OF AUTONOMY We can describe the four scenarios as follows: 1. I live an autonomous life which I have autonomously chosen. For example, a philosopher who has decided to live a life of constant questioning.

6 6 CRISPP 2. I live an autonomous life which I have not autonomously chosen. For example, an individual in a comprehensively liberal society which prohibits first-order non-autonomous forms of life such as military or convent life, or a child in a progressive school who is forced to think for herself and to question her surroundings. 3. I live a non-autonomous life which I have autonomously chosen. For example, an individual who, with no second-order autonomy-threatening social pressure, chooses to follow a restrictive religion or to join a convent or the army. 4. I live a non-autonomous life which I have not autonomously chosen. For example, an individual in a fundamentalist religious country, such as Afghanistan under the Taliban. Both political and comprehensive liberals will clearly endorse or at least allow case 1. Cases 2 and 3 are more difficult, and are directly opposed in some circumstances. For example, a woman who wants, under second-order autonomy, to enter a convent in which she has no first-order autonomy (case 3), will be unable to do so in a society which prohibits convent life on first-order autonomy grounds (case 2). In such a case, then, liberals who value autonomy will have to make a choice between first- and second-order autonomy: which is worthy of state protection? Nussbaum makes a number of statements in direct support of case 3, and which imply priority to 3 over 2, and thus priority to second-order autonomy rather than first-order autonomy. In A Plea for Difficulty, she defines the political liberal as someone who asks citizens to endorse a political conception of autonomy, one which carefully refrains from asserting that [first-order] non-autonomous lives are not worth leading, or even that autonomy is a key element in the best comprehensive view of human flourishing across the board; and she carefully protects the spaces within which Calvinists and other non-milleans [who do not value first-order autonomy] can plan lives according to their own lights. (Nussbaum 1999b: 110) It is important to note that, as I have highlighted in my insertions, Nussbaum is rejecting the universality of first-order and not second-order autonomy in this passage, although she does not make the distinction: her emphasis on protecting spaces for people to plan their lives according to their own lights shows that second-order autonomy remains a crucial universal goal. She stresses the universal value of second-order autonomy again when she states that political liberalism agrees with comprehensive

7 AUTONOMY, GENDER AND POLITICAL LIBERALISM 7 liberalism that a nonautonomous life should not be thrust upon someone by the luck of birth (ibid.: 110). This emphasis on second-order autonomy echoes the excerpt from Rawls quoted above: while (first-order) autonomy may not be a value for all people in all spheres of life, the question of whether a citizen has autonomy should still be decided by citizens severally, not imposed. Nussbaum s argument, then, is that political liberalism differs from comprehensive liberalism in preferring case 3 to case 2 when the two conflict, but that it agrees with comprehensive liberalism in disliking case 4: when a nonautonomous life is thrust on someone by luck of birth and so has not been chosen autonomously. For a political liberal understood in Nussbaum s terms, then, the crucial normative issue is not what kind of life a person lives in a first-order sense, but whether it has been chosen autonomously in a second-order sense. If it has, then the absence of firstorder autonomy is no cause for (political) concern. If a way of life has not been chosen autonomously, then there is cause for concern. If Nussbaum is correct in her description of the two positions, then, the debate between comprehensive and political liberals is not a debate about the value of autonomy per se, but a debate about the (political) value of first- vs second-order autonomy. 1 For political liberals, then, the state should provide a basic framework of freedoms. Within that framework, justice is served by giving individuals considerable leeway to construct their own conceptions of the good. Nussbaum captures this principle with what she and Amartya Sen term the capability approach : justice, and the governmental activity which promotes it, must focus on capabilities rather than on actual functioning. This focus on capabilities is, in turn, justified by an appeal to reason and the ability of individuals to choose their own ways of life. As Nussbaum puts it: for political purposes it is appropriate for us to shoot for capabilities, and those alone. Citizens must be left free to determine their course after that. The person with plenty of food may always choose to fast, but there is a great difference between fasting and starving, and it is this difference we wish to capture. (Nussbaum 1999a: 44) So, Nussbaum s commitment to autonomy is expressed through the political sphere. Individuals must be able to engage in practical reason about their way of life, choosing which ways of life to pursue and which to reject. Under such conditions, individuals may choose even those ways of life which are of no apparent value, or which do not enable autonomy

8 8 CRISPP to flourish. 2 In fact, Nussbaum implies that the very fact of choice imputes some worth to a way of life, by strongly distinguishing fasting and starving. On one level, there is very little difference: both have the same physical effects and are fatal at their extreme. The main difference between them is that one is chosen and the other imposed, and it seems to be this which makes the normative difference for Nussbaum. Choice becomes what I call a normative transformer, rendering an outcome just by its mere presence. Nussbaum does not appeal to the reason for fasting, or imply that the difference lies in the value of some ways of life above others. Thus she makes no distinction between the anorexic, the suffragette on hunger strike and the breatharian who has a spiritual belief in her ability to live on oxygen alone. All have chosen to fast. The state exhausts its duties by ensuring that all have the capability of eating, presumably by ensuring that no-one is unable to find or afford food. The fact that we might reasonably judge starving to death as bad for individuals, or as counter to their autonomy, does not justify state intervention. We protect individuals autonomy by protecting their ability to choose. If we retain Nussbaum s political liberalism, it becomes crucial to identify the conditions under which an individual comes to lead a way of life. For, even if an individual has autonomy on the [first-order] face of it, a political liberal might still want to interfere if it could be shown that comprehensive liberal meddling were to blame for the first-order autonomy. A political liberal would want to ensure that the individual could choose to alienate her own autonomy. 3 On the other hand, and more importantly for the purposes of this essay, a political liberal will excuse and protect a (first-order) nonautonomous life if and only if that life has been chosen autonomously. In order for Nussbaum s distinction between comprehensive and political liberalism to hold, it must be the secondorder autonomy status of a way of life which determines its susceptibility to or immunity from state intervention, and not the substantive first-order content. The Connection between First- and Second-order Autonomy But is it as simple as this? Consider case 3: the case where an individual has second- but not first-order autonomy. Imagine a woman who fits into box 3: Maria, who wants to be a nun. This second-order autonomous choice means that she will have to submit to the rules of the convent, and it is on that basis that I have labelled her as lacking first-order autonomy. But it might be objected that I have misdescribed the case. According to

9 AUTONOMY, GENDER AND POLITICAL LIBERALISM 9 this objection, what it means to be a nun is that one must submit to the rules of the convent. One cannot second-order autonomously choose to be a nun without, at the same time, autonomously choosing to submit to the rules of the convent. But, the objection continues, if one autonomously submits to the rules of the convent, one has first-order autonomy: in one s daily life, one is following rules that one actively wishes to follow, rather than following rules that are forced upon one or on which one does not reflect. It follows that Maria obeys the rules of the convent with firstorder autonomy (albeit without negative freedom). Without these firstorder desires, the second-order desire does not make sense. So, the objection concludes, the reason that political liberals defend case 3 and Maria s autonomous choice to be a nun is that her choice does not involve alienating her first-order autonomy at all. It realises that autonomy. We are thus left with two interpretations of convent life. According to the first interpretation, Maria has second-order autonomy (because she has chosen to become a nun) but not first-order autonomy (because, once she is a nun, she must obey the rules of the convent). According to the second interpretation, articulated in the previous paragraph, Maria has second-order autonomy (because she has chosen to become a nun) and first-order autonomy (because her choice to become a nun was in itself a choice to obey the rules of the convent, so that the convent rules are therefore followed autonomously). These alternative interpretations open up a number of complex issues, which will become clearer if we extend the Maria example. One rule of convent life (assume, for the sake of argument, that it is a requirement of all convents) is that nuns must wear habits. Assume that the important things about convent life for Maria are prayer and communal living, whereas rules of clothing carry no weight for her. Even on the first interpretation, once Maria is a nun she obeys the rules of prayer and communal life with both first- and second-order autonomy. Her second-order autonomy is realised because she wants to be a nun; her first-order autonomy is realised because she has considered, questioned and adopted wholeheartedly for herself the specific norms of prayer and communal life. What about Maria s submission to the rule that she must wear a habit? Clearly, if the habit were part of what Maria valued about convent life, if she had thought about the reasons for wearing a habit and found them separately motivating, she would have first-order autonomy regarding the habit even on the first interpretation. Ex hypothesis, however, this is not her situation. Maria wears the habit simply so as to serve her second-order goal of being a nun, despite finding it antiquated, uncomfortable and

10 10 CRISPP unrelated to her religious motivations and beliefs. Nonetheless, she chooses to wear it because, and to the extent that, it is a requirement of convent life. If there were no rule, she would not wear the habit. As there is a rule, she wears the habit so as to be able to stay in the convent. Does she, then, wear the habit with first-order autonomy? According to the second interpretation, which I suggested might be used to defend Nussbaum s position, Maria does have first-order autonomy. On this interpretation, it is not possible to break down the rules of the convent and say that they have different implications for Maria s autonomy. Wearing a habit is part of what it means to be a nun, and so it is not possible for her to will autonomously to be a nun without at the same time autonomously willing that she wear a habit. This answer, I suggested, might make the political liberal endorsement of this case, and other cases that would fit into box 3, more palatable. In effect, someone making this response would deny the very possibility of any box 3-like cases. 4 Anyone with second-order autonomy would thereby have firstorder autonomy with respect to all the norms that comprise the secondorder goal. Maria would, for this respondent, fit into box 1. This response, made from the context of the second interpretation, is problematic in two ways. First, it ignores the distinction between Maria s attitude to the habit and her attitude to prayer, and does so in a way that flatly contradicts the definition of first-order autonomy. First-order autonomy, I argued, is when an individual lives a daily life in which one is not beholden to but actively questions rules and norms, obeying only those that one endorses. If Maria s habit-wearing counts as an instance of first-order autonomy, it undermines the meaningfulness of this definition, of the distinction between the two types of autonomy, and of the distinction between the habit and prayer. 5 Second, even if we were prepared to undermine those distinctions and reject the possibility of considering the rules of convent life separately, the response is not one that Nussbaum or Nussbaumian political liberals can make. Remember, Nussbaum claims that political liberalism allows individuals to live nonautonomous lives, as long as those lives are uncoerced. Cases of type 4 (life under the Taliban) are ruled out, but cases of type 3 (Maria) are allowed. It is this feature that, according to Nussbaum, distinguishes political from comprehensive liberalism. But the hypothetical defence of Nussbaum we are considering effectively says that cases of type 3 do not exist. It moves all such cases into box 1, in which the individuals concerned have both second and first-order autonomy. In other words, it denies the very possibility of uncoerced (second-order autonomous)

11 AUTONOMY, GENDER AND POLITICAL LIBERALISM 11 nonautonomous (non first-order autonomous) lives. Such lives simply cannot exist, as a matter of conceptual impossibility. Nussbaum cannot follow this line of reasoning as to do so would mean the annihilation of political liberalism as she defines it. Nussbaum must, then, reject the idea that second-order autonomy entails first-order autonomy concerning the norms that comprise the overall goal. She must adopt the first interpretation and argue, correctly in my view, that Maria is not first-order autonomous as regards the habit. We might describe the situation as one in which wearing a habit serves Maria s second-order autonomy but not her first-order autonomy. Maria makes a second-order choice to forego her first-order autonomy as regards the habit. Maria wears the habit with second-order autonomy (because it furthers her goal of being a nun), but without first-order autonomy (because she does not endorse the rule for its own sake). Nussbaum s position is that such a situation is unproblematic from the point of view of justice. Maria lacks first-order autonomy but she does so from a position of second-order autonomy. She has made a choice to live a nonautonomous life which a political liberal must respect. I want to challenge Nussbaum s position, the prioritisation of secondorder autonomy, on two grounds. First, Nussbaum provides good reasons to be suspicious of the prioritisation of second-order autonomy (and thus to prefer the abandonment of political liberalism over the abandonment of the critique of female genital mutilation (FGM)) in her work on social construction or, as she calls it, the social formation of preferences (Nussbaum 1999a; 2000). Indeed, she suggests in this part of her work that opportunities for second-order autonomy are limited. However, there is a tension between Nussbaum s commitment to political liberalism and her concern to improve upon it: her political liberalism leads her to be wary of state intervention, and thus to prioritise an individual s ability to adhere to even those preferences which she has shown to be socially constructed and thus imperfect guides to justice. Once we have noticed that preferences are socially influenced, we can no longer maintain a position of political liberal non-intervention in the name of second-order autonomy. If preferences can be socially formed, then autonomy cannot require state non-interference on the basis that individuals must be left to make their own choices free from influence. Liberal theories of justice rest on two basic values: freedom or autonomy (understood in either a first- or a second-order sense) and equality. Nussbaum s political liberalism implies that the presence of second-order autonomy suffices to make a choice or way of life unproblematic from the

12 12 CRISPP standpoint of justice: such autonomy is a sufficient condition for justice. However, the social formation of preferences casts doubt on this position, in two ways. First, it suggests that people may be less autonomous than they appear, as their decisions are profoundly shaped by their social context. Second, if autonomy is (always) limited, a choice or outcome cannot be rendered just by the mere fact of having been autonomously chosen. We cannot determine whether a situation or practice is just by asking was it brought about solely by the autonomous choice of the individual concerned? for the answer will often (perhaps always) be no. Instead, the manner of their formation becomes a matter of justice. We must ask was the social influence which encouraged the individual to make that choice, and is the choice itself, compatible with justice? As Nussbaum denies that first-order autonomy is relevant to justice, we must turn to the second liberal value of equality. If individuals are subject to influence which threatens their equality then it is a requirement of justice that that influence be limited where possible. This intervention will take place not in the name of autonomy alone but in the name of justice, as a combination of autonomy and equality. Autonomy functions as a premise to the argument: if preferences have already been socially influenced, then protecting autonomy cannot simply be a matter of allowing individuals to follow their preferences, and preferences cannot determine the justice of a state of affairs. In order to distinguish between influence which does and which does not threaten justice, we must consider whether individuals are encouraged to make choices which threaten their equality. Second, then, I wish to argue that some ways of life or specific choices within a life are sufficiently problematic, even when chosen autonomously, as to merit state intervention. Nussbaum agrees with this position in the case of FGM, and this agreement causes problems: either she must extend that judgement to many other cases and reject her politically liberal prioritisation of second-order autonomy or, in maintaining her political liberalism, she must surrender her feminist critique of FGM and endorse a conservative state neutrality. This argument is extended in the next section. Second-order Autonomy and Harmful Norms So far, we have seen that Nussbaum, and political liberals in general, prioritise second-order autonomy over first-order autonomy. Put generally, the political liberal position is that, while a first-order nonautonomous life may not be thrust upon someone, it may be autonomously

13 AUTONOMY, GENDER AND POLITICAL LIBERALISM 13 chosen. But, as it stands, this position is too crude. It is too crude because people s choices take place in an unavoidably social context. People do not choose in a vacuum, with an infinitely wide variety of options and with ultimate control over their choices. Instead, people s choices are influenced, sometimes very strongly, by the people, structures and norms around them. As well as affecting people s preferences, social construction affects the options that are available. Social norms set out what may be chosen, and place conditions on what must be done in order to receive certain benefits. Moreover, social norms are sometimes harmful. If she is to follow a social norm, an individual may be required to harm herself. The harm may be physical, mental or material. It may also be social it may require the individual to lower her status relative to that of others. In other words, a social norm may reflect and perpetuate inequality. In some cases, a norm may be both harmful and unequal, in that it may require some people but not others to harm themselves. This second issue is the focus of the rest of this essay. It should become clearer if we consider another set of four examples. This time, instead of an intersection between first- and second-order autonomy, consider an intersection between second-order autonomy and the harmfulness or otherwise of social norms. Figure 2 represents possible answers to two questions. First, if an individual is following a social norm, has she chosen to do so with or without second-order autonomy? In other words, is she autonomously choosing to follow the social norm so as to acquire some benefit (including participating in a chosen way of life), or is she simply following the norm because it never occurs to her to do otherwise? Second, is the norm that she is following harmful? So as to simplify the example, assume that the individual in question has accurate and uncontentious knowledge as to whether the norm in question is harmful. 1. I am aware that there is a harmless social norm regulating access to some benefit. As I have autonomously chosen the benefit, I choose to follow the norm. For example, David is a member of an Oxford college. There is a norm operating in the college that people take tea in the common room at a certain time. Those who attend tea reap the benefits of forming social and professional networks. David considers the norm and, because he wants to further his academic career, decides to follow it by attending tea regularly I follow a harmless social norm regulating access to some benefit, but have not autonomously chosen to seek that benefit. For example,

14 14 CRISPP FIGURE 2 INTERSECTION BETWEEN SECOND-ORDER AUTONOMY AND HARMFULNESS OF SOCIAL NORMS David s colleague, Ben, also consciously attends tea so as to further his academic career. However, he has not autonomously chosen to be an academic: his career is the result of family pressure, his fear at the thought of leaving Oxford, and the fact that he has never got round to considering alternatives. 3. I am aware that there is a harmful social norm regulating access to some benefit. As I have autonomously chosen the benefit, I choose to follow the norm. For example, Rachel has autonomously chosen to seek career success. She works in a non-smoking office. She does not smoke, but her colleague and their mutual manager does. The colleague and the manager regularly go outside to smoke together, where they discuss business matters. As a result, the colleague is more included and favoured by the manager than is Rachel. In order to gain the manager s favour, and despite her dislike of smoking, Rachel decides to start smoking so that she can join the conversations in the smoking area and further her career I follow a harmful social norm that regulates access to some benefit, but have not autonomously chosen to seek that benefit. For example, in China, footbinding was necessary for a girl to secure a good marriage from the middle of the fourteenth century until the very start of the twentieth century. However, girls and women whose feet had been bound were left crippled and nearly housebound (Mackie 1996:

15 AUTONOMY, GENDER AND POLITICAL LIBERALISM ). Chun had her feet forcibly bound when she was a child. As a result, she was able to secure a good marriage, but suffered from severely limited mobility throughout her life. She made no autonomous choice to seek a life of marriage. INTERSECTION BETWEEN SECOND-ORDER AUTONOMY AND HARMFULNESS OF SOCIAL NORMS What are we to say about these examples? Which ought a liberal to endorse, and which ought she to criticise, or condemn as unjust? Most liberals would condemn case 4, the case of Chun and footbinding. The very similar case of FGM is discussed at length later in the paper. 8 Cases 1, 2 and 3 are more problematic, and highlight complexities with the political liberal prioritisation of second-order autonomy. If second-order autonomy is crucial to justice for political liberals, we might think that Ben, who has never made a choice to be an academic, is suffering from some sort of injustice. However, political liberals do not tend to be too worried about these sorts of cases because, I suggest, the life that Ben has ended up with allows him considerable first-order autonomy. Whilst we might think that his life would go better if he exercised his second-order autonomy and re-evaluated it, Ben does not seem to be a particularly needy victim of injustice. What of cases 1 and 3: David with his tea, and Rachel with her smoking? In both cases, Rachel and David are aware of the options that confront them, weigh up the costs and benefits of (non-)compliance with the norm, and choose to comply. But the costs that the norms impose are very different. Rachel will have to cultivate a taste for smoking which she knows will harm her quite considerably. David s tea, on the other hand, is certainly not harmful and has indeed been shown to have various health benefits. Does this difference, in and of itself, make a difference to the justice of the case? In the remainder of the essay, I argue that it does. Specifically, I argue that, in certain circumstances, individuals who submit to harmful norms so as to reach some higher, second-order goal are suffering from an injustice. Moreover, the liberal state ought to attempt to rectify that injustice. Perpetuating Inequality: Female Genital Mutilation and Breast Implants Generally, political liberals want to protect individuals second-order autonomy, even if it is used to choose ways of life that include practices which are harmful to the choosing individual. Mill s harm principle is perhaps the most famous example of such an approach (although Mill is not a political liberal), and I have already discussed Nussbaum s position.

16 16 CRISPP Sometimes, though, political liberals are willing to ban things for reasons of justice even when the injustice falls largely to those people who choose them. FGM causes problems for Nussbaum. As a feminist (and a Westerner? 9 ) she finds the practice abhorrent, and argues that it should be banned. As a political liberal, she lacks the philosophical resources to justify such a ban. Her arguments against FGM are effective arguments against political liberalism s prioritisation of second-order autonomy. Nussbaum s discussion of FGM gives many reasons for eradicating the practice. A consistent politically liberal approach to the issue would imply changing certain features of the practice, but leaving it optional for those adult women who wished to practise it as part of their conception of the good. However, not all of the criticisms that Nussbaum makes of FGM would be eliminated by that approach. In other words, FGM reveals that Nussbaum wants to label some practices as unjust regardless of whether they have been chosen autonomously, undermining her claim that, for a political liberal, second-order autonomy is sufficient for justice. Such practices are unjust, instead, because they threaten either the equality or the wellbeing of the choosing individual or, by contributing to social norms, they threaten the equality of a wider group of individuals. Nussbaum argues that we should consider FGM as worse than, and more worthy of elimination than, Western beauty practices such as dieting and cosmetic surgery. She gives eight reasons for her view (Nussbaum 1999a: 123 4): 1. FGM is carried out by force. 2. FGM is carried out on children below the age of consent. 3. Women who undergo FGM are more likely than Western women to be uneducated and thus to lack the conditions for autonomous choice. 4. FGM is often carried out in conditions which are dangerous to health. 5. FGM is irreversible. 6. FGM causes lifelong health problems. 7. FGM causes the loss of a certain type of sexual functioning that many women (would) value highly. 8. FGM is unambiguously linked to customs of male domination (Nussbaum 1999a: 124). Only the first three reasons are directly linked to the nature of the practice as unchosen. In other words, if second-order autonomy were the key factor in securing justice, FGM could be made acceptable if it were performed only on women above a certain age who gave their consent and who were given information about the risks involved and the particularity

17 AUTONOMY, GENDER AND POLITICAL LIBERALISM 17 of the custom the kind of information that is available to women who undergo cosmetic surgery in Western countries. 10 We could also eliminate some of the dangers to health (reason 4) by the provision of appropriate clinical equipment and training. These alterations would render the practice unproblematic from a strictly political liberal point of view (one which holds second-order autonomy as sufficient for justice, and which ignores the social formation of preferences). Are they enough? What is the force of the remaining objections? We are left with reasons 5 8. If we are political liberals of the kind Nussbaum wants us to be in A Plea for Difficulty, there seems to be nothing for us to say about these reasons that has relevance for action. FGM is irreversible (reason 5), but so are most tattoos, male circumcisions, abortions, precautionary mastectomies or hysterectomies and many sterilisations. The irreversibility of even a bodily procedure does not in itself suffice as a reason to ban it. FGM causes lifelong health problems (reason 6), but so do smoking, heavy drinking, unhealthy eating, inactive lifestyles and many other activities against which political liberals such as Nussbaum do not want to legislate. If we remove the forced element, then the loss of sexual functioning (reason 7) is insufficient for a ban remember, for Nussbaum, the person who has food may always choose to fast, and there seems to be no reason on the face of it why a political liberal should not allow an individual to deny herself something that others want for themselves. A political liberal does not forbid childless women in Western societies to have a hysterectomy to avoid unwanted pregnancy and the inconvenience of contraception and menstruation, 11 even though the capacity to bear children is one which many women value highly. Perhaps, then, objections 5 7 are not supposed to have any political implications they are reasons to dislike FGM, but nothing more. Nussbaum often does make a distinction between condemning practices and banning them, endorsing the former but not the latter, as long as the practice is autonomously chosen. 12 However, she does not make this distinction with FGM. In Women and Human Development Nussbaum argues that even consensual FGM (performed once objections 1 4 have been removed) could and should be banned, as it involves the permanent removal of a capability: it seems plausible for governments to ban female genital mutilation, even when practiced by adults without coercion: for, in addition to long-term health risks, the practice involves the permanent removal

18 18 CRISPP of the capability for most sexual pleasure, although individuals should of course be free to choose not to have sexual pleasure if they prefer not to. (Nussbaum 2000: 94) 13 This focus on capabilities is clearly an attempt to avoid the clash between political and comprehensive liberalism: Nussbaum wishes to show that FGM is non-negotiable because it falls within the jurisdiction of the basic capabilities on which all are supposed to agree. But, in order to remain a political liberal, Nussbaum frequently allows individuals to alienate themselves from certain of their capabilities. Sterilisation is an example. In true political liberal style, Nussbaum advocates state disinterest (Nussbaum 2000: 95). Nussbaum s political liberalism commits governments to ensuring that the basic capabilities are an option for all: that all are capable of the capabilities, as it were. However, she does not want to involve governments in forcibly ensuring that all individuals actually possess the relevant capabilities, regardless of their wishes on the matter except in the case of FGM. Her official, politically liberal position requires only that the conditions of capability are provided by the state. However, in considering FGM, Nussbaum reverts to the position which is best suited by her awareness of the social formation of preferences and the existence of harmful social norms: just because someone wants to do something, that doesn t necessarily mean that justice requires allowing them to do so. Nussbaum s position on capabilities does not, therefore, provide the philosophical resources to justify a ban on FGM tout court. Neither does her prioritisation of second-order autonomy. FGM can be thought of as an ingredient of a way of life that women may choose with second-order autonomy. Within cultures that practise it, FGM is often required for marriage. In such cases, we can think of FGM as an ingredient of marriage. More generally, we might think of FGM as a necessary ingredient of life within certain cultural communities: if women wish to maintain their membership, or to participate in the institution of marriage within it, they must undergo FGM. Some women might consent to undergo FGM, even though they might prefer not to, in order to become or remain eligible for community membership or marriage. Such women would be like Maria: they might choose their way of life (marriage or community membership) with second-order autonomy, and this might lead them to alienate their own first-order autonomy with respect to FGM. A powerful objection to FGM is thus that it is an excessively high price to ask women to pay in order to remain within their communities. But, as we have seen,

19 AUTONOMY, GENDER AND POLITICAL LIBERALISM 19 this is not an objection that can be made by Nussbaum if she is to prioritise second-order autonomy. For that prioritisation protects first-order nonautonomous lives i.e., those in which women have no choice but to undergo FGM if they wish to remain within their communities so long as they are chosen second-order autonomously i.e., so long as women are free to choose to leave their communities. It might seem at that Nussbaum can still justify a ban on FGM via second-order autonomy, in two ways. First, the option of leaving one s community is an extremely difficult one to take, and this fact undermines the idea that women do indeed undergo FGM as a result of a truly secondorder autonomous choice to stay within their communities. Second, it is the case that marriage is the only viable option for women in many cultures that practise FGM, so that women cannot really be said to choose marriage with second-order autonomy either. These points are strong. They point to the necessity of increasing women s second-order autonomy wherever possible, perhaps by trying to provide meaningful options for women other than marriage or community life. A political liberal prioritisation of second-order autonomy would demand such measures. It would not, however, demand a ban on FGM, for such a ban would increase the first- and not the second-order autonomy of women. A ban on FGM cannot, then, be explained by either the non-alienability of capabilities for that would require a ban on sterilisation or the prioritisation of second-order autonomy for that would require undermining the centrality of cultural membership and/or marriage. What, then, could explain the particularity of FGM? Reason 8 is one difference. Nussbaum writes: Female genital mutilation is unambiguously linked to customs of male domination. Even its official rationales, in terms of purity and propriety, point to aspects of sex hierarchy Sex relations constructed by the practice are relations in which intercourse becomes a vehicle for one-sided male pleasure rather than for mutuality of pleasure. By contrast, the ideal female body image purveyed in the American media has multiple and complex resonances, including those of male domination, but also including those of physical fitness, independence and boyish nonmaternity. These differences help explain why there is no serious campaign to make ads for diet programs, or the pictures of emaciated women in Vogue, illegal, whereas FGM is illegal in most of the countries in which it occurs. (Nussbaum 1999a: 124)

20 20 CRISPP This is indeed a salient difference between FGM and sterilisation. The former, and not the latter, results from and perpetuates forms of male domination. In other words, FGM undermines gender equality. Liberalism is based on two key values: freedom or autonomy, and equality. We have seen so far how Nussbaum incorporates autonomy into her political liberalism. With reason 8 against FGM, she introduces equality, specifically gender equality. However, the role of equality within political liberalism is somewhat ambiguous. Certainly, political liberalism requires that individuals are equal in the political sphere: that men and women have equal citizenship rights. What is unclear in the writings of both Nussbaum and Rawls, however, is the extent that equality in general and gender equality in particular may pervade other spheres. On the one hand, it is sometimes argued that gender inequality in religions, cultures and associations is acceptable so long as it is compatible with equal citizenship. Thus Rawls writes that the principles of justice do not apply directly to the internal life of churches and other associations (the examples he gives are universities and the family) so that, for example, ecclesiastical positions do not have to be allocated democratically or according to the difference principle (Rawls 2001: 164). He does not explicitly say whether church and other positions must comply with equality of opportunity; Nussbaum concludes that they must not and that gender discrimination is permitted. Thus she argues that political liberalism asks religions to accept the political equality of women as citizens, but claims this position is entirely compatible with allowing the Catholic church to employ only men as priests, for example (Nussbaum 1999b: 109). Since women are still able to vote, to leave the Church and to enjoy other rights of citizenship, gender discrimination is permissible in the priesthood. If the role of equality in political liberalism is limited in this way, then what is required to justify a ban on FGM is an argument to the effect that FGM prevents women from enjoying equal citizenship rights. Nussbaum provides no such argument; indeed, it would seem easier to argue in these terms against the male-only priesthood than against consensual FGM, for the former and not the latter has a direct bearing on women s ability to participate in the religious structures that determine the course of their own lives and the interface between those structures and the state. On the other hand, Rawls is at pains to point out that political liberalism does not abandon women to pervasive gender inequality in the family. He states categorically: It may be thought that the principles of justice do not apply to the family and that therefore they cannot secure equal justice for women and their children. This is a misconception (Rawls 2001: 163;